How MyFitnessPal Reimagined A New Era of AI-Enabled Nutrition Tracking with Tribe AI

Tribe

About MyFitnessPal:

MyFitnessPal is the #1 global nutrition and food tracking app for achieving health goals, especially weight management.

MyFitnessPal’s Challenge: 

As part of its mission to continuously improve the user experience, MyFitnessPal recognized the immense potential of incorporating GenAI within its product suite. MyFitnessPal (MFP) needed to bolster its internal team with an agency that had capacity to scope, prototype, and iterate on key innovative features.

Why Tribe AI? 

To overcome this challenge, the company brought in Tribe AI’s team of AI experts and practitioners to serve as an extension of MFP’s product organization, igniting its AI product roadmap and transforming its legacy products into AI-powered experiences that truly put customers first.

Proposed Solution:

Through two intensive four-week sprints, Tribe AI’s team leveraged AWS Bedrock and Anthropic Claude 2 models to transform MyFitnessPal's ideas into two functional Proof of Concepts(POCs):

(1) AI-powered voice logging, which allows users to effortlessly track their meals and nutrition by simply speaking into the app.

(2) A personalized recipe recommendation chatbot that provides tailored suggestions based on users' preferences and dietary needs.

Both features underwent internal testing before integrating them into MyFitnessPal's Premium-level product.

Tribe Team Members

Saguna: AI Product Manager (Voice Logging)

Faheem: AI Product Manager (Meal Planning)

Alex: AI Engineer

Gabriel: AI Engineer

Craig: Product & Engagement Lead

“Tribe has been a great partner and their model works exceptionally well. I’ve loved the flexibility and having a demo every week that we can share to the rest of the company.”
- Dustin Manderfeld, VP of Data Science & Engineering


Using Generative AI to Augment the MyFitnessPal User Experience


About MyFitnessPal:

MyFitnessPal is the #1 global nutrition and food tracking app founded in 2005 with a mission to help people around the world reach their health goals through better food choices by providing knowledge, motivation, and a sense of progress. Supporting nearly 1 million people in reaching their nutrition and fitness goals every year, and with a community of over 250 million users in 120 countries, MyFitnessPal offers members one of the world's most comprehensive nutrition and food tracking platforms, allowing them to track their food, record exercise activity, and log their weight. With one of the largest food databases in the world comprising over 19 million foods, access to over 2,000 recipes, more than 40 workout routines, and over 40 connected fitness partners, MyFitnessPal provides members with the tools for positive healthy change.

MyFitnessPal’s Challenge:
A daily staple of millions of health-conscious individuals, MyFitnessPal enables users to log their meals and nutrition intake by searching for foods and typing portions into the app.  Tracking food intake is a proven way to increase awareness of eating habits, supporting more informed decisions and healthier lifestyles. However, MyFitnessPal is always looking for ways to make things even faster and easier for their users.

To increase user engagement, MyFitnessPal sought to develop a voice-logging feature that allows users to log their food conversationally, with the system accurately and comprehensively recording items, quantities, and other relevant parameters in a single, streamlined step. Blending natural language processing with  GenAI, the system will translate speech to text, match it with food records in the database, and precisely record all user input, simplifying the food logging process for MyFitnessPal's users.

"I want to be able to hit a microphone button on the search bar and speak the two or three words of search terms so I don't have to type it,” said Dustin Manderfeld, MyFitnessPal’s VP of Data Science & Engineering. “If I can say everything I’ve eaten in a day, why can’t I do this all in one step?”

Why Tribe AI?
To bring this idea to life, Tribe AI and MyFitnessPal kicked off a four-week time-boxed prototype development process. After aligning the team on the project’s scope and expectations during the discovery week, Tribe AI used AWS Bedrock and Anthropic Claude 2 models to rapidly iterate and prototype the voice logging feature over the next two weeks. The rapid iteration process allowed the team to uncover additional capabilities, such as multi-language support helping them make informed decisions about which features to include in the final product.

Proposed Solution:
By the end of the sprint, the Tribe team of AI Engineers and Product Managers, built a functional demo capable of logging multiple items in a single interaction. The prototype, called Voice Log, makes food tracking as simple as having a conversation. Members can select the feature via the app's navigation bar, grant access to their phone's microphone if it's the first time they're using the feature, say what they ate in everyday language, and log the items based on the best matches suggested by the app. Members can review, make adjustments if needed, and tap "Log!" to complete the process.

Alternatively, users could give a partial query or item list, and receive prompts to input additional information, such as serving sizes, meal types, and brands.

Leveraging MyFitnessPal’s Expansive Food Database to Deliver Personalized Recipe Recommendations

MyFitnessPal’s Challenge:

Next, MyFitnessPal set out to answer a key question from their product team: Would users be willing to interact with a chatbot, either through voice commands or typing, to receive personalized recommendations and perform actions within the app? 

By leveraging a chatbot as an entry point for features like voice logging, updating weight, or navigating through different screens, MyFitnessPal aimed to make these processes quicker and more convenient, ultimately improving the user experience and increasing app engagement.

To address this, the Tribe AI team explored new ways of leveraging GenAI alongside MyFitnessPal's expansive food database (the largest of its kind!) to build a conversational chatbot. But with so many ideas for a ‘chatbot-based’ feature, the next four-week sprint with Tribe AI helped MyFitnessPal narrow its focus to one or two essential use cases.

"We had so many ideas in the arena of a chatbot and Tribe did a great job of challenging us to focus on one or two things and aligning the group. The biggest learning was that we can use generative AI to generate recipes."
- Dustin Manderfeld, VP of Data Science and Engineering

Proposed Solution
With a clear goal in mind, Alex and Gabriel again leveraged AWS Bedrock and Anthropic Claude 2, along with MyFitnessPal's unique internal data on users' food logging histories and preferences, to build a powerful recipe recommendation engine. This allowed the chatbot to provide personalized suggestions based not only on users' stated goals but also on their actual eating patterns, creating a differentiated and user-centric experience that went beyond the out-of-the-box capabilities of Claude 2 and offered a differentiated and powerful solution that truly caters to the needs of the app's users.

"There hasn't been a lot of confidence that our food database, because it's so vast and we have so many user-generated foods, would be able to integrate with those of other companies. But one great learning from the Tribe meal recommendation POC was that we can take these recipes, get the ingredients, and successfully map those through various techniques to our food database,” Manderfeld explained.

Tech Stack Details

Full-stack cloud-based application working alongside the existing MyFitnessPal’s environment: 

  • Cloud: AWS Bedrock
  • Large language models: Anthropic Claude 2
  • Languages used: 
    • Frontend: React
    • Backend: Django

What the Future Holds

In eight short weeks, TribeAI empowered MyFitnessPal to accelerate its AI product roadmap and unlock the full potential of its legacy products, augmenting and improving the user experience in ways that truly resonate with its customers. Tribe has also inspired the company's leadership to double down on a culture of experimentation and create a dedicated prototyping environment within MyFitnessPal.

With this agile approach and a commitment to continuous improvement, MyFitnessPal has positioned itself to quickly identify and develop innovative AI-powered features that will reinforce its position as a leader in the health and fitness industry. 

"I love Tribe’s method of committing to building and delivering a tangible product at the end of four weeks. This forces us to get on the same page and accelerate the product decisions we have to make. It's been a huge benefit." 

- Dustin Manderfeld, VP of Data Science & Engineering

Voice Log is currently available to all MyFitnessPal Premium iOs and Android members in the United States using the English language setting. For more information on Voice Log , visit blog.myfitnesspal.com or download the app for free via the App Store or Google Play.

Related Case Studies

Case Study

How Tribe AI Built a Model on GCP That Increased Security Questionnaire Auditor Efficiency by 55%

Case Study

Boomi Leverages Amazon Bedrock for Faster Help Desk Responses

Case Study

Insurance Company Uses ML to Optimize Pricing

Case Study

Accela Utilizes GenAI to Innovate 311 Help Lines with Faster & More Accurate Routing

Case Study

GenAI Solutions: How Bright Transformed Workforce Training with Tribe AI

Case Study

How Tribe Helped Reservoir Bring Finance Infrastructure to NFT Trading

Case Study

VitalSource Leans on GenAI to Reimagine Content Discoverability for Higher Ed Faculty

Case Study

Kettle uses machine learning to balance risk in a changing climate

Case Study

Native Instruments Leverages Amazon Bedrock for Smarter, More Intuitive Search and Discovery for Music Creators

Case Study

How Fantasmo is using machine learning to make GPS obsolete

Case Study

GoTo Revolutionizes Contact Center Quality Management with AI

Case Study

Tribe AI & rbMedia: Transforming Audiobook Production with Claude & Bedrock-Powered Dramatization

Case Study

Togal.ai powers the construction industry into the age of machine learning

Case Study

Taking a Data-Driven Approach to Women's Fertility with Rita

Case Study

Building a GenAI Roadmap for Educational Content Creation

Case Study

Building a Proprietary Investment Engine Using Public Data for a Top PE Firm

Case Study

Francisco Partners Accelerates Portfolio AI Efforts with Tribe AI

Case Study

Orchard Applies GenAI for a Faster, Easier-to-Use Lab Reporting Interface

Case Study

How Wingspan built a machine learning roadmap with Tribe AI

Case Study

How Togal AI Built the World's Fastest Estimation Software on AWS

Case Study

How Nota Built a Roadmap for AI-enabled Journalism with Help from Tribe

Case Study

How Tribe AI Shaped Truebit’s AI Strategy

Case Study

Tribe AI & Venture Labs: Accelerating Startups with Tailored AI Expertise

Tribe helps organizations rapidly deploy AI solutions that have real business impact.

Close
Tribe